One, and perhaps the only area of agreement in the business intelligence (BI) forum held recently was: BI is a complex field, so complex that nobody can even agree on what exactly the definition of BI is.
By Andrew Seldon, Editor, eSecure
Delving deeper into the topic is simply a journey into theory versus practice. Every person has their own experiences on what works and what does not, which must be tempered with the fact that customers often do not understand the real issues behind successful BI and delegate the project to IT or a strategic BI department, often to see the initiative suck up a lot of money without delivering results.
Sometimes setting up the model for BI is simple in concept, but when getting down to brass tacks, the moose hits the fan as different departments have informal processes and rules in addition to corporate rules which affect the model and the plans at each turn. Any implementation is also a political game, played with relish even though it costs more.
But there is much more that came out of the first Technews BI forum: Factors underpinning a successful BI initiative.
Technews and Systems Implementation Specialists (SIS) hosted the BI Round Table in October. We invited experts in the BI field to comment on what exactly makes BI implementation successful. The round table was such a success we decided to host them every second month in 2004 and report back in each issue of eSecure.
The BI round table is a forum, acting as a sounding board for the exchange of BI related viewpoints between like-minded individuals. It also acts as an incubator of ideas that can be utilised by its participants as well as the wider BI community. It is specifically a product, vendor, implementation and technology agnostic initiative with its primary objective being the discovery of best practices within the South African environment.
This article is the first of two covering the first round table. The topic, Factors Underpinning a Successful BI Initiative, was broken down into four segments:
3. Continued usage
4. Value derivation.
Each of our participants provided their experience to answer the questions. In this article, we cover some of the responses to design and implementation; more ideas, and some interesting comments from other experts will follow in the first issue next year.
The participants in the first round table were:
* Eugene Parsons, senior business intelligence consultant at Systems Implementation Specialists (SIS).
* Caron Mooney, director of IS Partners.
* Ian Teague, MD of NetMax Management.
* Michael Payne, manager: IT governance and security at Anglo Platinum.
* Peter Mills, business portfolio manager at T-Systems.
1. Eugene Parsons: Many BI initiatives fail because they are technological 'works of art' without much concern for business intelligence BI is a decision support system, especially at strategic and tactical levels. It is much more business than technology. These design phases ensure business' involvement from the outset:
a. Conceptual: identify problems/opportunities in need of decision support. Define metrics of performance against key strategic and tactical business objectives. This is the most difficult and crucial phase of a BI initiative and if done properly ensures success.
b. Logical: design the entire system on paper, an architectural blueprint which is crucial during system evolution.
c. Physical: implementation specific technology is incorporated into the design.
A process such as this provides for a frequent and iterative refinement process and solves specific decision support requirements.
2. Caron Mooney: The most important element of a successful BI implementation is a well-defined process or methodology that covers a number of aspects, including roles, deliverables, goals and risks. Business users and IT should drive the project together. Each on their own will not succeed.
In terms of design, a good technical methodology is required to ensure the right architecture is in place for the business. Subject areas must be well defined and the scope of the project tightly controlled.
If you are embarking on a BI project for the first time, it is strongly recommended that you engage with a technology consultant for at least part of the project. Nigel Pendse, author of the OLAP Report ( www.olapreport.com
) endorses this recommendation. One of the main reasons for this is that you need to tap into the experience of real-world implementations. There are many tips and best practices that only experienced BI implementers can effectively apply to each organisation's specific situation.
3. Ian Teague: In practical terms we link strategy and the measures of strategy to operations via management information system. Review of the drivers of the business is conducted in a way, which allows for revision of the thinking and value behind information and information flow. New thinking must be clarified, as it represents the basis of a new design.
The delivery mechanism involves mapping information flow via business functions and external sources into a system, which can solve information logistics and timing issues. Application of the system is bound within the practice of management by the decision makers of the organisation. In this aspect we review this practice and define the new potential for improvement in terms of a revised practice model.
4. Michael Payne: The bottom line with design is that the solution must solve the business problem. In arriving at that winning design, we need to understand what the user wants, and how we are going to build it.
1. Eugene Parsons: Rule number one: Ensure clean data. There is nothing worse than having empowered employees making correct decisions based on incorrect information.
A prominent and successful sponsor and driver of the project, both being business people, is essential to champion the BI initiative. At the same time the users must be involved as much as possible.
Most importantly, do not become obsessed with products. It is a great temptation to join the techies in their technological sandpit, however, keep your focus on the business issues of decision support and develop a corporate culture obsessed with knowledge rather than software.
Results are essential, therefore an incremental project-based rollout is necessary. Start small, bed down and proceed with the next phase of the project. This enhances the success rate by providing 'quick wins', each adding momentum and support for the BI system in its entirety. Soon the system is selling itself, which is not possible with a big bang approach.
2. Ian Teague: Plan to deliver measurable business and technical results. It is essential to consider the motivation and requirement of people to change and to focus on the competence and practice people require, working with the technology.
NetMax Management has learnt to work with people to bring about valued change.
3. Michael Payne: The implementation phase can be broken down into three phases:
The preparation phase is about understanding the environment into which this system is to be implemented. An organisation can be described as a melting pot of often conflicting ideas, beliefs, needs, wants, visions, values, strategies, insecurities, capabilities, competencies and political agendas, all operating over an IT infrastructure.